Catalytic Oxidative Ligand Transfer of Rh-Carbynoids and Alkyl Iodides

Abstract

Herein, we disclose the first catalytic generation of alkyl–I^(III) Rh^(II)-carbynoids via oxidative ligand transfer between alkyl iodides and aryl–I^(III) Rh^(II)-carbynoids. When homoallylic iodides were used, alkyl–I^(III) Rh^(II)-carbynoids evolved through a diastereoselective cyclopropanation that led to highly electrophilic bicyclic alkyl–I^(III) species. We observed that the latter species reacted with a plethora of nucleophiles, enabling access to a valuable class of cyclopropanes that were converted to housanes.

Rh^(II)-carbynoids, a class of Rh^(II)-carbenes substituted with a leaving group at the carbene carbon atom, have emerged as key species in the catalytic transfer of monovalent cationic carbon units (:⁺C–R) (Figure A). Our group pioneered their catalytic generation using diazo hypervalent iodine reagents, and we demonstrated that our Rh^(II)-carbynoids can behave as a Rh^(II)-carbene and undergo [2 + 1] cycloadditions with alkenes and alkynes. ,, The latter processes led to cyclopropyl–I^(III) and cyclopropenyl–I^(III) that ultimately evolved to allyl and cyclopropenium cations, respectively. On the other hand, we also observed that the Rh^(II)-carbynoids could exhibit the reactivity of hypervalent iodine species and undergo nucleophilic attack by carboxylic acids to give a Fischer-type acyloxy Rh^(II)-carbene or para-selective electrophilic aromatic substitution to generate a donor/acceptor Rh^(II)-carbene. DFT calculations carried out for the reaction with carboxylic acids underpinned an initial attack of the carboxylic acid to the hypervalent iodine^(III).

1.

Oxidative ligand exchange of Rh-carbynoids with alkyl iodides.

Such behavior prompted us to question whether our Rh^(II)-carbynoids could undergo other classes of activations characteristic of hypervalent iodine^(III). An interesting process in this area is the oxidative ligand transfer between an aryl–I^(III)X₂ and an alkyl–I^(I), which produces an alkyl–I^(III)X₂ and aryl–I^(I). Taking into account that our Rh^(II)-carbynoids are functionalized with an aryl–I^(III) moiety, we wondered whether such oxidative ligand transfer could occur with readily available alkyl–I^(I) compounds. If successful, this process would catalytically generate alkyl–I^(III) Rh^(II)-carbynoids, which remain inaccessible since the corresponding hypervalent iodine reagents are synthetically inaccessible. Moreover, we anticipated that alkyl–I^(III) Rh-carbynoids could participate in [2 + 1] cycloadditions with unsaturated substrates and generate novel and unexplored classes of alkyl–I^(III).

Here, we report the successful development of such a concept for the catalytic generation of alkyl–I^(III) Rh-carbynoids. Using alkyl iodides substituted with an alkenyl group in a remote position, the novel Rh-carbynoid underwent an intramolecular diastereoselective cyclopropanation that led to a distinct class of bicyclic alkyl–I^(III) intermediates. The latter species were unstable above −20 °C and, consistent with their electrophilic nature, underwent regioselective nucleophilic attack by a broad range of nucleophiles. We found that the iodocyclopropane products obtained had utility for the construction of housanes.

Originally, we proposed a mechanism where alkyl iodide 1a would undergo an oxidative ligand exchange with an aryl–I^(III) Rh^(II)-carbynoid int-I , forming alkyl–I^(III) Rh^(II)-carbynoid int-II . The latter species would evolve through intramolecular cyclopropanation, leading to bicyclic hypervalent iodine int-III . Finally, the formation of cyclopropanes 3 would take place from an α- or β-attack by suitable nucleophiles (Figure ).

2.

Mechanistic hypothesis.

The feasibility of the proposed catalytic generation of alkyl–I^(III) Rh-carbynoids was initially explored at −50 °C by stirring alkyl iodide 1a with Du Bois catalyst Rh₂(esp)₂ in dichloromethane, followed by slow addition of aryliodine^(III) diazo reagent 2a and subsequent addition of tributylmethyl­phosphonium dimethylphosphate as the nucleophile. To our delight, iodocyclopropane 3a was obtained in 58% yield with excellent diastereoselectivity (dr > 20:1, Table , entry 1) by selective α-attack of the nucleophile to the corresponding int-III . We did not observe formation of allylic phosphate 3a* resulting from the direct cyclopropanation of int-I and 1a followed by electrocyclic ring opening. ,

1. Optimization Studies .

entry	2	1a:2	Rh cat.	T (°C)	yield 3a–c (%)	ratio 3:3*
1	2a	1:1.2	Rh2(esp)2	–50	58	>20:1
2	2b	1:1.2	Rh2(esp)2	–50	50	10:1
3	2c	1:1.2	Rh2(esp)2	–50	40	>20:1
4	2a	1:1.2	Rh2(OAc)4	–50	n.d.	-
5	2a	1:1.2	Rh2(TFA)4	–50	n.d.	-
6	2a	1:1.2	Rh2(TPA)4	–50	n.d.	-
7	2a	1:1.2	Rh2(oct)4	–50	13	>20:1
8	2a	1:1.2	Rh2(adc)4	–50	50	5:1
9	2a	1:1.2	Rh2(OPiv)4	–50	35	18:1
10	2a	1:1.2	Rh2(esp)2	–40	61	>20:1
11	2a	1.2:1	Rh2(esp)2	–40	72	>20:1
12	2d	1.2:1	Rh2(esp)2	–40	65	>20:1
13	2e	1.2:1	Rh2(esp)2	–40	67	>20:1

^aReactions performed at 0.1 mmol scale. Yields reported on the basis of ¹H NMR analysis of the crude reaction using CH₂Br₂ as internal standard.

^b 3:3* ratios were determined by ¹H NMR analysis of the crude. esp = α,α,α′,α′-tetramethy-1,3-benzenedipropanoate. TFA = trifluoroacetate. TPA = triphenylacetate. Oct = octanoate. Adc = adamantane-1-carboxylate. OPiv = pivaloate.

Modulation of the ester substitution in reagent 2 to benzyl (2b) or trichloroethyl (2c) affected the reaction outcome: the former enhanced the formation of the carbon-insertion product 3b* (E/Z > 20:1), while the latter slightly reduced the yield (Table , entries 2, 3). Alternative rhodium paddlewheel catalysts, including Rh₂(OAc)₄, the more electrophilic Rh₂(TFA)₄, Rh₂(TPA)₄, and Rh₂(oct)₄, failed to catalyze this transformation (Table , entries 4–7). In contrast, the more sterically demanding Rh₂(adc)₄ and Rh₂(OPiv)₄ delivered 3a, albeit in diminished yields (Table , entries 8, 9), suggesting that steric and electronic tuning of the catalyst substantially impacts the efficiency of the oxidative ligand transfer process. Further optimization involved elevating the reaction temperature to −40 °C and adjusting the stoichiometry between 1a and 2a, both of which improved the reaction outcome (Table , entries 10, 11). Finally, we observed that both pseudocyclic and linear triflate reagents led to 3a in slightly lower yields (entries 12, 13).

After the optimization studies, we explored a broad range of heteroatomic nucleophiles and observed that negatively charged nucleophiles derived from tetrabutylammonium salts of nitrate (3d), mesylate (3e), triflate (3f), halides (3g–j), cyanide (3k), azide (3l), or benzoate (3m) provided the corresponding cyclopropanes from moderate to good yields (Table A). Neutral heteroatomic nucleophiles such as alcohols (3n), ethers (3o), thiol (3p), secondary and tertiary amines (3q–s), cyclic and linear amides (3u,v), and nitrogen-heterocycles (3w) performed well. On the other hand, activated arenes such as anisole (3x) or mesytilene (3y) as well as allyltrimethylsilane (3z) provided the corresponding C–C bond-forming products. It is worth highlighting that, in any case, we did not observe even traces of the corresponding β-attack products.

2. Nucleophile and Alkyl Iodide Scope for the Synthesis of Cyclopropanes 3 .

^aReactions performed at 0.2 mmol scale using 1.2 equiv of the corresponding alkyl iodide and 1.0 equiv of reagent 2.

^bAllylic iodide byproduct 3ap* was isolated in 20% yield.

^cLinear hypervalent iodine reagents Ph­(OTf)­I–R 2f (R = COPh), 2g (R = CO₂CH₂Cl₃), and 2h (R = CF₃) were used. Yields are reported on the basis of the isolated pure product using flash column chromatography. Diastereomeric ratios are >20:1 unless otherwise stated in brackets and were determined by ¹H NMR analysis of the crude reaction. Relative configurations of 3a, 3am, and 3ao were assigned based on ¹H–¹H NOESY experiments and confirmed for 3w by single-crystal X-ray diffraction analysis.

After this, we decided to evaluate the alkyl iodide scope using trans-styryl derivatives substituted at the para (3aa–af), meta (3ag–ah), and ortho (3ai) positions as well as naphthalene derivatives (3aj,ak) and heterocycles (3al), which were well tolerated (Table B). The methodology proved amenable to trisubstituted olefins, as exemplified by the efficient formation of 3am with excellent diastereoselectivity. 1,1-Disubstituted terminal alkenes were also tolerated (3an), highlighting the method’s capacity to accommodate differently hindered alkene substitution patterns. The use of cis-styryl homoallylic iodide (Z/E > 20:1) led to the formation of 3ao as an 8:1 mixture of diastereoisomers, suggesting that the cyclopropanation step likely proceeds via a cationic intermediate susceptible to isomerization. Notably, 1,3-diene provided access to 3ap, indicating selective reaction at the homoallylic double bond while preserving the distal olefin for potential functionalization. In addition, single-carbon insertion byproduct 3ap* was also isolated. Nonactivated aliphatic olefins derived from 2-pentenylbenzene and octene delivered 3aq and 3ar in excellent yields, and alcohol-protected aliphatic alkenes were also well tolerated (3as,at). Moreover, our process enabled the synthesis of spirocyclic compound 3au and bicyclic product 3av, thus highlighting the versatility of our protocol in reaching complex cyclopropanes.

After this, we investigated the influence of the iodoalkyl chain length of 1 (Table C). trans-β-Iodostyrene and cinnamyl iodide failed to deliver the expected products 3aw and 3ax, likely due to the high ring strain associated with the formation of the corresponding three- and four-membered bicyclic alkyl–I^(III) intermediates ( int-III ). However, extending the carbon chain by two methylene units (n = 3) led to the formation of 3ay in 50% yield. Further elongation of the alkyl chain resulted in a loss of reactivity with traces of detectable formation of the desired product 3az.

The effect of substitution in the alkyl chain was examined by introducing a methyl group in the allylic position of the homoallylic iodide, furnishing 3ba in 55% yield as an equimolecular mixture of diastereoisomers (Table C). In contrast, methyl substitution in the same carbon bearing the iodine led to traces of 3bb, likely due to steric clashes between the methyl group and Rh^(II)-carbynoid. Finally, we were delighted to observe that alternative ketone, ester, or trifluoromethyl linear reagents 2f–h were well tolerated (Table D, 3bc–be).

Encouraged by the success of the scope evaluation, we next wondered whether alkyl–I^(III) Rh-carbynoids could be generated from simple alkyl iodides such as methyl iodide and evolve through intermolecular alkene cyclopropanation. We observed that iodocyclopropanes 4a,b could be obtained from styrenes albeit in low efficiency and low or no diastereocontrol. However, cyclic alkenes such as cyclohexene or 1,4-cyclohexadiene provided cyclopropanes 4c and 4d in good yields and diastereoselectivity (Figure A).

3.

Intermolecular oxidative ligand exchange, detection of 3a-int-III , and control experiments.

On the other hand, successful attempts to detect the corresponding bicyclic hypervalent iodine int-III intermediates (see Figure ) were carried out with 1a and 2a under the optimized reaction conditions in CD₂Cl₂. We observed the clean formation of 3a- int-III at −40 °C, whose structure was assigned by 1D/2D NMR experiments and HRMS. Although we observed that 3a- int-III was stable up to −20 °C, we failed to confirm its structure by X-ray diffraction analysis (Figure B). Encouraged by these results, we focused on detecting by ¹H NMR intermediate 4c- int-III produced from cyclohexene, reagent 2a, and methyl iodide. However, only formation of 4c was observed, which suggested a lower stability of 4c- int-III in comparison to 3a- int-III (Figure C).

Finally, a reaction carried out with chiral catalyst Rh₂(S-NTTL)₄ (under the previously optimized reaction conditions for the enantioselective single-carbon insertion into alkenes) with homoallylic iodide 1a and 2b led to a separable mixture of cyclopropane 2b and allylic phosphate 2b* (Figure D). The disparity in enantiocontrol in the formation of both products clearly supported two different reaction mechanisms. In this sense, the formation of int-III by alkene cyclopropanation with int-I and subsequent oxidative ligand exchange is less likely than the proposed mechanism depicted in Figure .

After the development of our oxidative ligand transfer of alkyl iodides with Rh-carbynoids, we questioned the synthetic potential of the iodocyclopropane 3 products. In 2024, the group of Marek reported a stereocontrolled synthesis of bicyclo[1.1.0]­butanes (BCBs) using iodocyclopropanes containing a leaving group in an appropriate position. The cyclization was promoted by the generation of a lithium cyclopropyl intermediate with nBuLi via Li–I exchange. Inspired by this work, we thought that our cyclopropyl derivatives 3 could be suitable starting materials for the synthesis of bicyclo[2.1.0]­pentanes, usually named housanes. Such strained, sp ³-rich scaffolds remain underexplored in medicinal chemistry due to lack of general methodologies. This is in sharp contrast with other small-ring systems such as cyclopropanes, cyclobutanes, or bicyclo[1.1.1]­pentanes (BCPs) known to improve pharmacokinetic properties, target selectivity, and clinical success rates.

To prove our hypothesis, we treated bromide derivative 3i and nBuLi in THF at −78 °C and observed the desired housane product 5a (35% yield) and cyclopropane 5* (50% yield), which was generated from the protonation of the corresponding cyclopropyllithium intermediate (Figure A, entry 1). Then we observed that while tBuLi provided poor yields, LDA (Li–NiPr₂) led to 5a as the major compound (Figure a, entries 2, 3). As expected, the nature of the leaving group significantly impacted the reaction outcome, observing the highest yields for the iodocyclopropyl derivative 3j, and no formation of 5* took place (Figure A, entries 4–7).

4.

^a Reactions performed at 0.1 mmol with 3a, 3f, 3h, 3l, 3j, and 2.2 equiv of lithium base in THF at −78 °C for 1 h; yields are reported on the basis of ¹H NMR analysis of the crude reaction using CH₂Br₂ as the internal standard. ^b Reactions performed with 3j, 3ad–ak, 3am, 3an, 3aq, 3ar, 3ay, 3au, 3bc, 3bd (0.1 mmol, 1.0 equiv), LDA (0.22 mmol, 2.2 equiv), and THF (2.0 mL) at −78 °C for 1 h; yields are reported on the basis of the isolated pure product using flash column chromatography. The relative configuration of 5 was assigned by analogy of that of 7, which was determined by single-crystal X-ray diffraction analysis. Diastereomeric ratios are >20:1 and were determined by ¹H NMR analysis of the crude reaction.

After identifying iodide as the suitable leaving group and LDA as the lithium–halogen exchange agent, we were delighted to observe that a series of iodocyclopropane derivatives substituted with different aromatic and aliphatic groups were well tolerated, observing formation of housanes 5b–p in good yields and excellent diastereocontrol. Moreover, under the optimized reaction conditions, compound 3ay with a one-carbon longer iodoalkyl chain was successfully transformed in bicyclo[3.1.0]­hexane 6 in good yield (Figure C). Finally, ester hydrolysis of 5a under basic conditions provided housane carboxylic acid 7, whose structure was confirmed by single-crystal X-ray diffraction analysis (Figure D).

In conclusion, we have discovered a new catalytic activation of alkyl iodides with aryl–I^(III) Rh-carbynoids that led to alkyl–I^(III) Rh-carbynoids. The latter species evolved through diastereoselective inter- and intramolecular cyclopropanations to produce linear and cyclic alkyl–I^(III). While we were unable to detect linear alkyl–I^(III) species at low temperatures, we characterized cyclic derivative 3a- int-III produced from a homoallylic iodide. Such cyclic alkyl–I^(III) species were generated from a broad range of homoallylic iodides and derivatized with negatively charged and neutral heteroatomic/carbon nucleophiles to produce iodocyclopropanes with excellent diastereoselectivity. The latter served as precursors of functionalized housanes by using a lithium–iodide exchange and intramolecular alkylation. Current work focuses on the search and design of a dirhodium catalyst that provides access to cyclopropanes 3 with excellent enantiocontrol while suppressing the direct single-carbon insertion into the alkene.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c09559.

• Experimental procedures and spectral data (PDF)

The authors declare no competing financial interest.